Replication data for:

Lachapelle, Jean, Steven Levitsky, Lucan A. Way, and Adam E. Casey. 2020. "Social Revolution and Authoritarian Durability." World Politics 72, no. 4. doi:10.1017/S0043887120000106.

# Files description 

## revregimes.csv
Dataset of revolutionary regimes.

## revregimes_tscs.csv
Dataset of revolutionary regimes in time-series cross-sectional format.

## codebook - revolutionary autocracies.pdf
Codebook for dataset.

## raw_ts.csv
Full dataset for the analysis.

## analysis_main.R
Replication of the main analysis. 
*Input*: main_data.Rdata and synth.Rdata

## analysis_appendix.R
Replication of the analysis presented in the appendix. 
*Input*: main_data.Rdata and synth.rData.

## prepare_datasets_mi.R
Multiple imputation and variable transformations. 
*Input*: raw_ts.csv and *ruling coalitions in dictatorships, 1946-2008.csv*. 
*Output:* main_data.rData.

## synth_control_meth.R
Synthetic control analysis. 
*Input*: raw_ts.csv. 
*Output*: synth.Rdata.

## functions.R
Useful functions for survival analysis and the synthetic control method.

## pool_compare.R
Functions for performing likelihood ratio tests for nested Cox models with multiple imputation.

## ruling coalitions in dictatorships, 1946-2008.csv
Dataset of ruling coalitions by Milan Svolik (2012) 

# Variable list:
- *Country*: country name
- *ccode*: COW country code
- *cus_caseid*: regime identifier
- *cus_syear*: regime start year
- *cus_smonth*: regime start month
- *cus_sday*: regime start day
- *cus_eyear*: regime end year
- *cus_emonth*: regime end month
- *cus_eday*: regime end day
- *cus_fail*: regime failure, binary indicator
- *cus_sr*: revolutionary regime (binary)
- *regimetype*: authoritarian regime type
- *cus_AR*: authoritarian regime (binary)
- *cus_t_surv*: total duration of regime
- *cus_t*: regime age
- *partB*: regime has a party component (binary)
- *mil*: regime is pure military (binary)
- *persB*: regime has a personalist component (binary)
- *mon*: regime is a monarchy (binary)
- *prev_partB*: previous regime was a party-based authoritarian regime (binary)
- *prev_mil*: previous regime was pure military  (binary)
- *prev_persB*: previous regime was personalist (binary)
- *prev_mon*: previous regime was a monarchy (binary)
- *prev_demo*: previous regime was a democracy (binary)
- *cus_ARevnum*: cumulative count of previous authoritarian regime breakdowns
- *sv_comm*: communist regime (based on Svolik 2012)
- *InstalledCom*: foreign installed communist regime
- *e_migdppcln* : GDP gap per capita (log) from Maddison (2018), as provided by V-Dem (2019)
- *e_migdpgro*: GDP growth from Maddison, as provided by V-Dem (2019)
- *tpop*: total population
- *ln.tpop*: population (logged)
- *oil_gas_valuePOP_2014*: per capita value of oil and gas production
- *ln.oil_gas_valuePOP_2014*: log(oil_gas_valuePOP_2014 + 1)
- *e_regionpol*: geographic region, as provided by V-Dem (2019)
- *milper*: military personnel
- *milperpc*: military personnel per capita
- *ln.milperpc*: log(milperpc)
- *war*: interstate war (binary)
- *v2xcs_ccsi*: Core Civil Society Index (CCSI), from V-Dem (2019)
- *v2xel_frefair*: Clean Election Index, from V-Dem (2019)
- *reg.civsoc.lnd*: regional average of Core Civil Society Index (CCSI) among countries that share a land border
- *polity2*: Polity IV score
- *coup1*: first coup attempt for a given year. Coded as 2 if coup is successful, 1 if coup fails, and NA if there was no coup attempt. Source: Powell and Thyne (2011) (v2017.11.28)
- *coup2*: second coup attempt for a given year
- *coup3*: third coup attempt for a given year
- *coup4*: fourth coup attempt for a given year
- *date1*: date of first coup attempt
- *date2*: date of second coup attempt
- *date3*: date of third coup attempt
- *date4*: date of fourth coup attempt
- *partymilit*: level of party control over the military. Source: Geddes, Wright and Frantz (2018)

## Notation
The *L* suffix after a variable means lagged. For instance *tpopL1* refers to population lagged one year. *tpopL2* refers to population lagged two years, etc. Similarly with *Lead*: *tpopLead1* denotes population size with a one year lead.

## Variables for alternative datasets of authoritarian regime spells 

### Geddes Wright and Frantz (2014):
- *gwf_caseid*: regime identifier for GWF dataset (same as GWF_casename)
- *gwf_AR*: authoritarian regime (binary)
- *gwf_sr*: revolutionary regime (binary)
- *gwf_spell*: regime duration
- *gwf_fail*: regime failure (binary)
- *gwf_t*: regime age

### Svolik (2012):
- *svol_caseid*: regime identifier for Svolik dataset
- *svol_AR*: authoritarian regime (binary)
- *svol_sr*: revolutionary regime (binary)
- *svol_fail*: regime failure (binary)
- *svol_t_surv*: regime duration
- *svol_t*: regime age

### Magaloni, Chu and Min (2013):
- *mcm_caseid*: regime identifier for MCM dataset 
- *mcm_AR*: authoritarian regime (binary)
- *mcm_sr*: revolutionary regime (binary 
- *mcm_fail*: regime failure (binary)
- *mcm_t_surv*: regime duration
- *mcm_t*: regime age





